Despite the success of current anti-HIV drugs, which can often reduce levels of virus in a patient to undetectable levels, the drugs do not cure people of HIV, so that virus levels rapidly rebound if a patient stops taking the drugs. The major reason for this failure is that low levels of latent or reservoir HIV remain despite the drug treatment, hiding out in long-lived and quiescent cells such as central memory T cells. Strategies that removed this residual reservoir could complement current antiretroviral therapies and allow for the eventually cure of infected individuals. Here, we describe an approach to reduce the latent reservoir through the targeted delivery of HIV-specific nucleases to a key subset of cells that has been shown to be enriched in latent HIV genomes. The approach uses emerging technologies based on targeted nucleases and engineered viral vectors, and will be evaluated using in vivo models of HIV latency. In this way we propose to develop a new class of therapeutics that could contribute to the management of HIV infection and the goal of a cure.